Polymerase chain reaction (PCR) amplification of portions of the genome of both rapidly growing mycobacteria and nocardiae, followed by restriction fragment length polymorphism (RFLP) analysis of the amplification products, has proven to be a useful technique in the diagnostic laboratory. Identification of many isolates to the species level can be obtained within a few days of organism isolation using this technique, as compared with the month or more required for conventional identification based on biochemical testing. In addition, these molecular procedures allow better discrimination among species and subspecies than is possible with biochemical testing, and facilitate the detection of hitherto undescribed species. Our work with two different areas of the Nocardia genome (a portion of the gene for 16S ribosomal RNA and a portion of the gene for the heat-shock protein) has suggested the existence of several such unrecognized Nocardia species; work is ongoing to characterize these organisms further. In addition, we have found several clinical isolates belonging to the species Nocardia veterana, newly recognized both as a species and as a human pathogen. A manuscript describing has just been published which describes our findings with regard to this organism, including the problems with distinguishing it from another newly described pathogenic species, Nocardia africana. Some isolates of Nocardia species cannot be precisely identified even with our RFLP procedure or with complete 16S rDNA sequencing. DNA-DNA hybridization is currently being utilized to define these isolates more precisely. Characterization of several such isolates, based based both on phenotypic and molecular biologic features, is currently underway. We also hope to characterize further several interesting Nocardia isolates we have found that appear to possess several different 16S rRNA gene copies per cell. It would be of taxonomic, biochemical, and medical interest to determine the extent of similarity among these genes, and which of them is functional. In addition, we have begun collaborations with others who have extensive organism collections to assess the extent to which some of their isolates may belong to species that can only reliably be identified using molecular methods.